Zebrafish embryonic and adult neutropoiesis display differential requirements for Gfi1 paralogs.

Abstract

Neutropoiesis describes the developmental process that generates neutrophil granulocytes, white blood cells that play essential roles in innate and adaptive immunity. In zebrafish embryos, neutrophils form in three overlapping waves. The first two waves generate primitive neutrophils in the anterior and posterior lateral mesoderm (ALM and PLM). These cells are derived from haemangioblasts, bipotent mesenchymal progenitors of blood and endothelial cells. Adult-type neutrophils form in the third wave from hemogenic endothelial cells (HECs) of the dorsal aorta. Unlike zebrafish, mice generate no primitive neutrophils. All their neutrophils develop from HECs. Their formation requires the transcriptional repressor Gfi1. Gfi1 mutant mice suffer from severe neutropenia and die soon after birth. The zebrafish genome encodes two Gfi1 proteins, Gfi1aa and Gfi1ab. Here, I show that there is a differential requirement for Gfi1aa and Gfi1ab during the different waves of zebrafish neutropoiesis. The ALM expresses Gfi1ab, yet the formation of ALM-derived primitive neutrophils was unaffected in gfi1ab mutant embryos. The PLM expresses Gfi1aa but the loss of Gfi1aa expression led to an increase rather than a loss of PLM-derived neutrophils. The same increase was also observed in gfi1aa/1ab double mutants. A single-cell RNA sequencing experiment showed that the expanded double mutant primitive neutrophils were able to mature but displayed markedly altered gene expression profiles. Increased expression of anti-inflammatory genes like il4 and tim2b correlated with a delayed neutrophil response to a sterile wound injury in gfi1aa single and gfi1aa/1ab double mutant embryos. In the double mutant larvae, the initial increase in primitive neutrophils was followed by a complete loss of neutrophils at later stages, suggesting a defect in the third wave of neutropoiesis. While the double mutants failed to thrive, larvae that retained a single copy of either gfi1aa or gfi1ab were viable, yet their kidney marrow neutrophils were not necessarily normal. Adult fish with a single copy of gfi1aa carry kidney marrow neutrophils with abnormal light scatter and Sudan black staining characteristics. My data suggest that insufficient levels of Gfi1 protein, rather than qualitative differences between the Gfi1aa and Gfi1ab proteins, cause the phenotypic abnormalities. In summary, this project provides new insights into the diverse roles of Gfi1 proteins in different waves and stages of zebrafish neutropoiesis and raises new questions that open new avenues of investigation.